of water ice and hydrated minerals (1, 5, 12).
Consequently, we simulated changes in the hydration state of regolith materials by adding or
removing hydrogen in the form of water to or
from model compositions. This allows hydrogen
concentration to be expressed in terms of water-equivalent hydrogen (WEH). The regolith was
modeled as a two-layer structure, with an ice-free upper layer covering icy soil (Fig. 2, inset),
consistent with ice-stability models ( 13). Both
layers have the same non-icy composition, with
the lower layer containing an additional, variable fraction of water ice.
Figure 2 shows simulated hydration trends for
selected materials with different Fe concentra-
tions. Their arched shape results from an initial
increase in the low-energy neutron population
with small amounts of added hydrogen, which
enhances capture gamma-ray production, fol-
lowed by a decrease in gamma production as
Fe is diluted by water. With layering, the model
counts deviate from the hydration trends, fol-
lowing different paths between nonlayered end
members as depth is varied. For each material,
there is a corresponding set of counts spanned by
permutations of hydration and layering.
For comparison with the models, Ceres meas-
urements were normalized to globally averaged
counting rates for Vesta. The vestan measure-
ments were further normalized to a model in
which Vesta’s global regolith composition was
assumed to be similar to the howardite meteor-
ites with the addition of 250 mg/g H, the lower
bound determined by GRaND at Vesta ( 11). In
this case, Ceres data plot inside the layered set
for the CI and CM model compositions (Fig. 2).
Furthermore, the data points form a concave
up pattern, consistent with trends for layering.
If Vesta’s regolith contained more H, the data
points for Ceres would shift toward the hy-
dration trends for the CI and CM chondrites.
However, the lower bound is our best estimate,
given that most of the H was delivered to Vesta
by carbonaceous chondrite impactors, as detailed
in ( 10).
Near the equator and within GRaND’s field
of view, stability models predict that water ice
is present at depths greater than those sensed
by GRaND (Fig. 3, B and C). Consequently, hydration trends describe the equatorial measurements (e.g., yellow points in Fig. 2). These
measurements do not follow a specific hydration
curve, which indicates that [Fe] is variable at the
equator (square brackets denote concentration).
Because the hydration trends for CI and CM
chondrites do not intersect the data points, they
can be excluded as representative compositions.
For Ceres, the maximum equatorial points are
consistent with simulated material E1 ( 10), which
is poorer in Fe than the average CI chondrite.
Thermal + epithermal
Fe 7. 6 MeV
0E 180E 360E
0E 180E 360E
Fig. 1. Maps of corrected neutron and gamma-ray counting rates from data acquired at low altitude. (A) 6Li(n,a) reaction rate in lithium-loaded glass (thermal
plus epithermal neutrons). (B) Full energy interaction rate of 7.6-MeV Fe capture gamma rays in bismuth germanate. East longitude convention is used ( 17), and data are
superimposed on shaded relief. The white circle in (A) indicates the approximate full width at half maximum (FWHM) spatial resolution at a selected equatorial
measurement location. The white lines indicate the FWHM at the poles. The absolute rate and dynamic range are shown at right, along with rates averaged over latitude
bands (zonal averages) and average pixel uncertainties (1s error bars). The dynamic range for the same measurements at Vesta is shown for comparison (V).